A dimensional exploration and scale development study of algorithm aversion

The increased use of AI in business has spurred an explosion in algorithm aversion research. The absence of scientific measurement instruments has caused the empirical research on the structural dimensions and measurement scales of algorithm aversion to stagnate, and the field is currently just in the exploratory stages of investigation. The results of experimental research and polls on algorithm aversion and appreciation may not be as broadly applicable as they may be because roughly two thirds of them used U.S. samples. Thus, extending from previous research, this work applies grounded theory to investigate the dimensionality of the structural dimensions of algorithm aversion using data from Chinese user interviews as well as the MicroBlog, Zhihu, and CSDN corpus. The scale was tested through the processes of questionnaire, exploratory factor analysis, and validation factor analysis to construct the scale of algorithmic aversion. The study finds five dimensions of algorithm aversion: Algorithm power gameplay, Algorithm user lock-in, Algorithm cognitive bias, Algorithm recommendation preference, and Recommendation algorithm adoption. The scale has a good level of validity and reliability and comprises 22 items. The findings of this study will support theoretical underpinnings for AI marketing and practical research on algorithm aversion in recommendation systems.